The combined effects of extracorporeal membrane oxygenation and renal replacement therapy on meropenem pharmacokinetics: a matched cohort study

Extracorporeal membrane oxygenation (ECMO) is being increasingly used in adult patients with acute severe cardiorespiratory failure as a supportive therapy [1],[2]. Prolonged support for bridge to recovery or transplantation is now possible. While ECMO sustains life, stabilises physiology and allows time for definitive management, little is known about the independent effects of ECMO on antibiotic pharmacokinetics (PK). ECMO is thought to further complicate the PK alterations seen during critical illness [3], which appears to manifest as increased volume of distribution (Vd) and decreased clearance (CL) [4]. Ex vivo, animal [5],[6] and clinical studies [7] are currently underway to further investigate the PK changes seen during ECMO and to develop evidence-based dosing guidelines. Simulated ex-vivo studies that utilised adult circuitry [8] have demonstrated significant antibiotic drug sequestration in the circuit based on physicochemical properties of individual drugs. However, the PK data on antibiotics in critically ill adult patients is limited with available studies indicating significant PK alterations [3],[9],[10]. This is concerning as the risks of suboptimal drug dosing (both under- and overdosing) are profound in this complex group of patients who have high infection-related mortality.

A significant number of patients receive ECMO for severe cardiac and/or respiratory failure resulting from infectious aetiologies. Patients may develop new infection during ECMO support. Studies indicate that infections occur frequently during ECMO and infections/colonisation with multi-drug-resistant organisms is not uncommon [11]-[13]. Gram-negative bacteria are responsible for significant proportions of these infections acquired during ECMO [11]. Meropenem is used as an empirical or targeted broad-spectrum antibiotic in this setting. It is a minimally protein-bound and hydrophilic drug that undergoes significant sequestration/degradation in ex vivo ECMO circuit models [8]. In this setting, one would anticipate profound alterations in meropenem PK in patients on ECMO, although to date, we are unaware of any studies to guide meropenem dosing in adults on ECMO.

This open-label, descriptive, matched-cohort PK study aimed to describe single-dose meropenem PK during ECMO using critically ill patients with sepsis and not receiving ECMO as controls.

Five (two VV, three VA) out of eleven ECMO patients received RRT. The other six patients (four VV, two VA) had no significant impairment in renal and hepatic functions, based on routine biochemical parameters. The indications for ECMO included pneumonia, septic shock (n = 7); cardiogenic shock (n = 2); sickle-cell crisis (n = 1); primary graft dysfunction post lung transplant (n = 1). The median sequential organ failure assessment scores were not significantly different between the controls and ECMO patients (7 [3]-[15] vs. 13 [9]-[15],[17],[18], respectively, P = 0.14). The demographic and clinical data are summarised in Table 1. Median time to PK sampling in ECMO patients was 2 days (1 to 7).

This study provides preliminary evidence that standard meropenem dosing (1 g IV 8-hourly) as an intermittent bolus infusion in ECMO patients is likely to result in drug concentrations sufficient to treat highly susceptible Gram-negative pathogens. Conventional-dose meropenem should achieve a time over minimal inhibitory concentration (T_>MIC) of 100%, assuming a minimum inhibitory concentration (MIC) of 2 mg/L (the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for Pseudomonas aeruginosa). However, when treating less susceptible P. aeruginosa (MIC₉₀ 8 mg/L) and Acinetobacter species (MIC₉₀ 16 mg/L) higher meropenem doses would have to be considered especially in patients with elevated CrCL. Given that, patients on ECMO have decreased CL in most cases [3], standard dosing is likely to achieve target plasma concentrations in most patients. This is important considering the potential clinical and bacteriological benefits of maintaining 100% T_>MIC in critically ill patients [19].

This study uses meropenem plasma concentration data from four different patient populations to perform robust dosing simulations and to provide preliminary insights into the incremental effects of critical illness, ECMO and RRT on meropenem PK. The plasma concentrations observed in ECMO patients reflect a balance between the independent alterations in Vd and CL that occur in the presence of critical illness [20], organ failures and ECMO [3]. Interestingly in this study, routinely targeted meropenem plasma concentrations (>2 mg/L) were maintained with standard dosing, both in ECMO patients on RRT and those with preserved renal function. However, plasma meropenem concentrations were significantly higher in the RRT group when compared to patients with preserved renal function. This is important as standard dose adjustments for renal impairment (for example IV 500 mg 8-hourly or 1 g 12-hourly) in these patients receiving RRT may potentially result in under dosing. Equally, use of higher than standard doses may precipitate the risk of toxicity. Therapeutic drug monitoring where available may further improve the safety and efficacy of meropenem dosing during ECMO [4],[21].

ECMO patients demonstrated reduced meropenem CL and an increased Vd when compared with controls, but these changes were not statistically significant. This trend is consistent with the PK changes expected during ECMO based on available literature [3]. An increase in Vd resulting from critical illness [20] and sequestration in the ECMO circuit [8] can significantly affect plasma concentrations probably of meropenem, a hydrophilic with limited protein binding and predominant renal CL. Equally, AKI is common in patients on ECMO, with incidence as high as 70% to 85% in single-centre studies [22]. The Extracorporeal Life Support Organisation (ELSO) Registry data [23] suggests that up to 46% of patients on VV ECMO and 44% on VA ECMO may require some form of RRT during the ECMO run. There is significant variability in mode of RRT used in ECMO patients and this may appear to limit the generalisability of our results. This to an extent has been overcome with our dosing simulations that account for a range (20 to 180 mL/min) of net CrCL (native and RRT) achieved in ECMO patients However, dosing may not be entirely based on CL achieved during RRT and possible residual renal CL or extra renal CL may have to be considered. This is of high relevance during ECMO as meropenem can undergo significant degradation/sequestration during their transit through the ECMO circuit. Although upregulated non-renal elimination is possible for ciprofloxacin in renally impaired patients [24], there is no data to support this in the case of meropenem.

In this study the estimated median meropenem CL seen in controls on CVVHF was 3.5 (3 to 4) L/h. There is no reliable data on meropenem CL during EDD-f even in non-ECMO patients and it is highly likely that this will be greater than seen with high-volume CVVHF. Despite RRT partially compensating for decreased drug CL in patients with ECMO and acute kidney injury (AKI), they maintained significantly higher meropenem concentrations during the entire dosing interval with standard dosing when compared with patients without RRT and the controls. Our simulations confirm that a meropenem dose of 500 mg - 1 g 8-hourly will provide a plasma concentration >2 mg/mL in 90% of the patients with CrCL ranging from 20 to 180 mL/min. Given the relatively wide therapeutic index of meropenem, highly variable CrCL between ECMO patients based on modality and intensity of RRT used, loss in the ECMO circuit and preponderance of less susceptible organisms in this population especially with prolonged ECMO support, a dose of 1 g 8-hourly may be considered appropriate till more PK data becomes available. Meropenem accumulation and under dosing are still potential concerns in patients with extremes of CL and these high-risk groups need to be specifically addressed in future PK studies in this population.

The findings of this study contradict the available sparse data pertaining to meropenem PK during ECMO. To our knowledge, there are no previously published PK studies in neonatal or adult patients on ECMO. A recent case report [9] indicated heightened meropenem CL (20.8 L/h) and Vd (0.56 L/kg) during ECMO and RRT and a high-dose meropenem infusion was utilised to maintain optimal concentrations. However, the CL for meropenem in the current study was significantly lower (7.3 ± 5.6 L/h) and the Vd was highly comparable (0.53 ± 0.17 L/kg). However, it should be noted that meropenem is unstable at 37°C and ongoing exteriorization of blood during ECMO may lead to a degree of spontaneous degradation, which can be erroneously interpreted as increased CL. While it is challenging to arrive at any strong conclusions based on these data, it appears that eventual success of meropenem regimens during ECMO may rely more on the CL that occurs in an individual patient.

There is emerging data to suggest that the commonly used dose of meropenem (1 g 8-hourly) may be sufficient to treat an unselected population of septic critically ill patients not receiving ECMO or RRT [25]. In this setting, the risk of under dosing with 1 g 8-hourly dosing in critically ill patients on ECMO and with preserved renal function appears small and augmented renal clearance [26] has not yet been described in this population. However, patients on peripheral VA ECMO in whom oxygenated blood is being returned into iliac artery or distal aorta may experience very high non-pulsatile renal blood flows and whether this result in heightened CL in patients with preserved renal function needs further evaluation.

The current study has limitations. Characterizing altered PK in patients receiving RRT while on ECMO can be complex. Variability in the techniques used for RRT and ECMO is a significant limitation in the generalisability of our results. Future studies should further investigate the effects of type and intensity of RRT chosen on meropenem PK during ECMO. Despite the ECMO population being small and heterogeneous, our model could accurately predict drug concentrations in ECMO patients and controls and discriminated well for RRT. This study does not address the pharmacodynamic aspects of meropenem therapy and no meaningful clinical outcome measures can be generated from the small sample. The non-ECMO patients selected were considered to be optimal controls for patients on ECMO who exhibited systemic inflammatory syndrome with or without clear evidence of infection. Systemic inflammation is known to significantly affect volume of distribution of the hydrophilic and renally excreted drugs such as meropenem [20]. Hepatic and renal function, however, were not matched as a clear separation between controls and ECMO patients who had preserved renal function and ECMO patients who were dialysis dependent, desirable in the context of this study, which sought to highlight the influence of inflammation and illness, ECMO and RRT.

In patients receiving meropenem on ECMO, standard dosing (1 g 8-hourly) should achieve routinely targeted plasma concentrations. However, an increase in dose may be necessary when targeting higher plasma concentrations (>4x MIC and 100% T_>MIC) and or in patients with elevated creatinine clearance. Future PK studies should validate these findings especially in ECMO patients with extremes of CrCL (<20 or >150 mL/min). Therapeutic drug monitoring where possible is recommended until robust dosing guidelines become available.